Estimation of human health risk from exposure to methylmercury via fish consumption in Ghana

Fish is an important source of nutrition in many parts of the world. A fish consumption advisory enables countries such as the United States to recommend the amount of specific species of fish that may be eaten based on estimates of possible health effects associated with contaminants. These advisories may include recommendations to limit or avoid eating certain species of fish caught from specific water bodies or from water-body types (e.g., lakes, rivers or coastal waters) due to chemical contamination.1


Introduction
Fish is an important source of nutrition in many parts of the world.A fish consumption advisory enables countries such as the United States to recommend the amount of specific species of fish that may be eaten based on estimates of possible health effects associated with contaminants.These advisories may include recommendations to limit or avoid eating certain species of fish caught from specific water bodies or from water-body types (e.g., lakes, rivers or coastal waters) due to chemical contamination. 1 In Ghana, fish is a primary source of protein, due to its low cost compared to other animal proteins and because of its beneficial nutritional factors such as healthy fat content. The nutritional benefits of fish, especially the supply of long chain fatty acids, have been noted by the World Health Organization (WHO). 4This resource, however, may be affected by many domestic and industrial activities that introduce pollutants to the water.The issue of pollution has become a matter of concern, especially because some of these pollutants, notably metals, bioaccumulate in fish such that concentrations found in fish tissue are higher than those found in the surrounding water.Studies conducted in the Faroe Islands and New Zealand suggest that high consumption of contaminated fish may lead to harmful levels of exposure to metals, although there is complexity in interpretation and generalization of these findings to other populations.Metal pollution has been a major international issue, especially for nonessential metals like methylmercury (MeHg), cadmium (Cd) and lead (Pb).MeHg is a neurodevelopmental toxin. 8The epidemiologic evidence of the effects of metals are epitomized by the case of methylmercury (MeHg) poisoning in Minamata, Japan in the 1950s, when a chemical plant contaminated the water, resulting in widespread cases of nervous system disorders and deaths from the consumption of contaminated fish and seafood. 9In addition, loss of cognitive and motor skills in children exposed to methylmercury in utero via fish and seafood (including whale and shark meat) consumption has been identified in the Faroe Islands and New Zealand.The occurrence and severity of the effects of exposure to methylmercury, cadmium and lead depend upon the amount to which a person is exposed and characteristics of the individual, including life stage.

Literature Search
Data were gathered from scientific publications identified using PubMed, a web-based database that comprises over 23 million citations for biomedical literature. 19The following Research kg), using the average (25 kg) of the FAO estimation of fish consumption in Ghana of 20 to 30 kg per year per capita. 25,26We then multiplied this value by the mean concentration of mercury that was found in fish in published works identified in our literature search. The JECFA health guideline value, which is the provisional tolerable weekly intake (PTWI) of 0.0016 mg/kg bw, was divided by 7 days to attain a per day value of 0.0002 mg/kg bw/day. 29lso, the RfD of 0.0001 mg/kg/day from the USEPA IRIS for methylmercury was applied to the HQ calculation for comparison. 30e cancer risk from eating contaminated fish is usually expressed using the life time average daily dose by the slope factor (SF), which is a 95% upper bound confidence interval.However, for methylmercury, the SF was not derived, because available data shows that systemic non-cancer effects would be seen at methylmercury exposures lower than those required for tumor formation. 30

Statistical Analysis
The mean concentration of mercury in each reported species resulted in 63 samples in our analyses.Mercury concentrations in unidentified species of fish were included in the initial analysis (n=4), but were not included in the trophic level analysis. 31The distribution of concentrations was not normal, therefore non-parametric analyses were employed.
One-way analysis of variance (ANOVA) was used to compare the means of the groupings (source of fish and trophic levels).We used the Kruskal-Wallis test to compare mercury concentrations among the different sources of fish (water type) as well as the trophic level of the fish.A Mann-Whitney test was carried out to compare the differences between the auriferous vs. other regions.
Significance levels for all tests were set at an alpha=0.05.All statistical analyses were performed using GraphPad Prism version 5.00 for Windows.

Overview of Hg Database
Based on the literature search, mercury was the most common metal pollutant evaluated in fish in Ghana.The mercury contents in fish were analyzed by the use of cold vapor atomic fluorescence spectroscopy or an open flask procedure developed at the National Institute for Minamata Disease (NIMD) in Japan by Akagi and Nishimura (1991). 32The recovery of referenced material or spiked samples was over 95%.The detection limit, when provided in articles, was 0.5 ng Hg g-1.Sixty-three samples comprising 53 different species were analyzed from 10 different water sources.Data were not sufficient to further evaluate other metals of interest, since only one study in the searched database had results for cadmium and lead.All of the identified fish species were species commonly consumed by Ghanaians.Marine sources provided 36.8% of the different fish species included in the analysis, followed by lakes/reservoirs at 33.3%, and rivers accounted for 29.8% of fish species included in the analyses.All of the studies were conducted in the southern half of Ghana, with the exception of one study in the Upper Volta Basin in the Yeji, Brong-Ahafo Region of Ghana (Figure 1).The highest value of total mercury detected was 0.89 mg/kg ww in the Kafue pike fish (Appendix 1).The trophic levels of the fish were also determined. 21-23Fish species within trophic levels 2-3, 3-4, and 4-5 were considered low, middle and high trophic levels respectively.The locations of the water bodies where the fish were caught were mapped and identified as a river, lake/reservoir, or marine body of water.

Non-cancer risk assessment
Using the map generated by Hilson, the samples were divided into auriferous (gold-bearing) vs. other regions. 24

Calculation of health risks
To calculate non-cancer health risks, the HQ was used.The HQ is a widely used estimate in risk characterization and is the ratio of the estimated exposure to a chemical over the level at which no adverse effect is expected.
The HQ is derived by dividing dietary intake (DI) by the health benchmark, hence an HQ of less than 1 means no adverse health effects are expected and an HQ greater than 1 means adverse health effects are possible. 14To derive the DI, we calculated the average fish consumption per person per day (0.068 Doke, Gohlke quotient (HQ) was used.Table 1 presents the values used to determine daily intake, exposure duration, and ultimately the hazard quotient.Based on values presented in Table 1, the estimated safe level of methylmercury exposure for an adult male Ghanaian was calculated as 0.013 mg/day (0.0002 mg/kg bw/day * 65kg) using the JECFA PTWI and 0.0065 mg/day (0.0001 mg/kg bw/day * 65kg) using the USEPA RfD.
Of the fish samples identified in the studies, 13% of the fish had mercury concentration levels that produced an HQ of more than 1 using the JECFA PTWI.All of the samples that produced an HQ of > 1 were from river sources.In comparison, using the more conservative USEPA standards, 33% of the fish had an HQ > 1.Out of these sites with an HQ > 1, 52% were from rivers, 38% were from coastal marine waters and 10% were from lakes/reservoirs.The remaining sites (67%) had an HQ < 1, meaning that adverse health outcomes are unlikely for an adult male or female, based on the concentration and consumption rate of fish.It is important to note the HQ calculations are based on estimated average daily exposures over a lifetime.
Table 2 presents the mercury levels and HQ associated with the source of the fish sampled.Out of the sixty-three (63) samples, the mean concentration of mercury in fish from rivers, lakes/ reservoirs, and marine areas were 0.25 (±0.23), 0.04 (±0.04), and 0.06 (±0.04) mg/kg ww, respectively.Significant differences were identified between the mercury concentration among the fish from the various water sources (p <0.0001), with significant differences between mercury in fish from rivers compared to marine areas and lakes/ reservoirs.All other differences in sources of mercury in fish tested were not significant at p < 0.05 (Figure 2).Mean concentrations of mercury were highest in fish from riverine sources in the auriferous regions of Ghana (mean concentration of 0.25 mg/kg ww). 20nfortunately, there were no samples taken from rivers in non-auriferous regions to allow a direct comparison.Samples from Lake Bosumtwi and Lake Volta (Yeji), had the lowest mercury levels (mean of 0.04 mg/kg ww).When the mercury database was sorted into samples from auriferous regions vs. samples from non-auriferous regions, a Mann-Whitney U test showed that there were highly significant differences between the mean of the samples taken from the auriferous regions (0.22 mg/kg ww) vs. the non-auriferous regions (0.05 mg/kg

Table 1 -Parameters used in estimating health risk a,b Body weight measurements were back calculated from BMI values from a study of 4377 participants in the Greater Accra Region. 43 BMI is given as weight in kilograms divided by the square of height in meters (kg/m2). 44 * Provisional Tolerable Weekly Intake (PTWI) Table 2 -Summary of mean concentrations of mercury/HQ per type of water body
Estimation of human health risk from exposure to methylmercury via fish consumption in Ghana

Type of Sample Water Body
Doke, Gohlke ww).Fish from the auriferous region produced an HQ > 1 irrespective of whether WHO or USEPA standards are used, and those of the non-auriferous region produced an HQ < 1.
A comparison of the trophic levels of fish showed significant differences between the mean mercury concentrations of high (0.14±0.20) and low trophic levels (0.03±0.01), as well as between the middle trophic level (0.09±0.11) and the low trophic level.However the high and middle trophic levels were not significantly different, meaning that the mercury concentration was not dependent on whether fish were in the high or middle trophic group (Figure 3).The World Health Organization and United States Department of Agriculture recommend approximately 2 servings of fish per week, whereas current intake estimates in Ghana suggest approximately 5 fish servings per week. 25,26,29,33-35Assuming only 2 servings per week (weekly serving size of 0.17 kg) and a mean mercury concentration of 0.10 mg/kg, the associated HQ is < 1 even using the USEPA more stringent RfD of 0.0001 mg/kg/bw per day.However, in individual fish samples, one sample out of the 63 was associated with an HQ > 1 even at the recommended twice weekly serving using the JECFA PTWI.This sample was of Hepsetus odoe (Kafue pike), taken from River Pra.Using the USEPA RfD, the recommended 2 servings per week generates an HQ > 1 in 10% of the samples.All of these samples were from river sources (Appendix 1).
A summary of all the datasets compiled indicates an average mercury concentration of 0.10 mg/kg ww ± 0.15 in fish from Ghana.The estimated overall HQ is less than 1 using the Joint FAO/WHO Expert Committee on Food Additives Provisional Tolerable Weekly Intake (JECFA PTWI), and slightly above 1 (1.1) using the USEPA RfD.The difference between HQ values for males (0.5 (± 0.8)) vs. females (0.6 (± 0.8)) using JECFA estimates and also males (1.1(± 1.6)) vs. females (1.2 (± 1.7)) using the USEPA RfD was negligible (Appendix 1).

Discussion
However, it is also important to develop and evaluate risk management and communication strategies that both minimize risks and maximize benefits from fish consumption. 4In order to enhance the benefits of fish consumption, generation and analysis of monitoring data on contamination levels is critical.
This study provides an overall summary of the mercury contamination status and possible health risks associated with methylmercury in fish in Ghana.When considering a mean mercury concentration of 0.10 mg/kg wet weight in fish, our results indicate showed no significant differences however the comparison between lakes/reservoirs vs. rivers, as well as marine areas vs. rivers yielded significant differences at p < 0.05.
Doke, Gohlke methylmercury is not likely to cause health effects at the estimated mean fish consumption levels of approximately 5 servings per week, using the JECFA PTWI.This initial attempt to estimate the potential risk associated with methylmercury in fish reveals River Pra to be the more polluted water body in terms of mercury contamination (Appendix 1).In total, the most polluted fish samples were from rivers in the gold mining region, with a mean Hg concentration of 0.25 mg/kg ww, corresponding to an HQ > 1 using both the JECFA and USEPA health values.The least polluted were the reservoirs with a mean of 0.03 mg/kg ww, corresponding to an HQ < 1. Concentrations of methylmercury were noted in the order: rivers> marine> reservoirs/lakes.
Our results suggest that 12.6% of the fish analyzed in Ghana are above the level at which adverse health effects are possible using the JECFA PTWI, and 33% using the more conservative USEPA RfD.These adverse health effects are possible from a lifetime consumption of contaminated fish at the estimated consumption rate of 68 g/ day.The mean mercury concentrations found in the present study (0.10 mg/ kg ww (± 0.15)) in fish in Ghana are lower than the mean levels found in a recent study by Karimi et al.That study compiled mercury concentrations in fish that may be consumed by the US population.The mean level of Hg in fish (calculated from their supplemental table) was found to be 0.29 mg/kg ww (± 0.55). 38is study is the first paper to synthesize the available data on methylmercury exposure via fish consumption in Ghana.However, the study is limited by several factors.There was a limited data set (10 publications) and the geographical spread of the samples did not include northern regions of Ghana.It is also important to note that both the USEPA RfD and JECFA PTWI include uncertainty factors, thus the use of the HQ is meant as a conservative estimate of the potential for human health effects and is not measuring risk directly.Limitations of the epidemiological studies used in the estimation of the dose-response relationships are also important to note and have been critiqued previously. 39 also note that fish is usually eaten cooked in Ghana Further study into the cooking methods of Ghanaians and the methylmercury concentrations in cooked fish is needed to estimate a more accurate level of exposure.

Conclusion
The fish collected and analyzed for Hg in Ghana thus far have a relatively low average level of MeHg (~0.1 µg/g), although higher levels are seen in rivers in the gold mining regions (0.25 (±0.23)).These results suggest that regular monitoring of fish collected from water bodies in gold mining regions is warranted.Results also suggest that regulatory bodies may want to consider development of guidelines for fish consumption advisories when warranted, and remediation of primary sources of mercury contamination to optimize the health benefits of fish consumption.
Estimation of human health risk from exposure to methylmercury via fish consumption in Ghana

Comparisons of mercury content by the trophic levels of the fish showed significant differences between the high vs. low trophic level (p-value = 0.02) and middle vs. low trophic level interaction, but no significant differences were seen in the high vs. middle trophic level interaction.
Doke, Gohlke

Figure 1 -
Figure 1 -Sampling sites for mercury in fish in Ghana.Sampled sites with yellow diamonds indicate lakes/ reservoirs, blue squares indicate marine areas and red circles indicate rivers.

Figure 2 -
Figure 2 -Mercury concentration in fish samples (mean ± SEM) from lakes/ reservoirs, marine, and rivers in Ghana.The blue line indicates the level 0.19 mg/kg, which represents the mercury concentration corresponding to an HQ of 1 using the JECFA reference dose.The red line indicates the level 0.095 mg/kg, which represents an HQ of 1 using the USEPA reference dose.This is the maximal concentration limit in fish, considering the dietary intake rates in Ghana, and the reference dose of methylmercury below which no adverse effects are likely.Comparing mercury concentrations in fish from lakes/reservoirs vs. marine areasshowed no significant differences however the comparison between lakes/reservoirs vs. rivers, as well as marine areas vs. rivers yielded significant differences at p < 0.05.

Figure 3 -
Figure 3 -Mercury concentration in fish by trophic levels (mean ± SEM).Comparisons of mercury content by the trophic levels of the fish showed significant differences between the high vs. low trophic level (p-value = 0.02) and middle vs. low trophic level interaction, but no significant differences were seen in the high vs. middle trophic level interaction.

5-7
. A study by Obodai et al. showed that Hg concentrations in cooked S. melanotheron (0.334 mg/kg ±0.14) samples were significantly lower than the concentrations recorded in uncooked S. melanotheron (0.479 mg/ kg ±0.13). 40However, some regulatory bodies like the United States Food and Drug Administration and the California Office of Environmental Health Hazard Assessment do not consider cooking fish to significantly reduce the amount of methylmercury in the fish. 41,42